The treatment of locally advanced rectal cancer combines chemoradiation followed by radical surgery. A number of rectal cancer patients treated with preoperative chemoradiation have no detectable cancer cells in the surgical specimens. Our long-term objective is to determine whether or not patients with a complete pathologic tumor response to chemoradiation can achieve local control without radical surgery. The benefit to such patients in terms of reduced morbidity and improvement of quality of life, as well as the marked reduction in health care costs will be significant. The potential use of chemoradiation as the only form of therapy in rectal cancer patients is appealing, yet the information on rectal cancer response to radiation is very limited. The reported rate of pathologic complete pathologic response varies between series due to differences in the chemoradiation protocol. Whether pathologic complete response represents transient tumor regression or permanent eradication remains unknown. The diagnostic methods used to image rectal cancer cannot accurately differentiate preoperatively between radiation-induced fibrosis and residual tumor;pathologic assessment of the surgical specimen is the only way to diagnose pathologic complete response. In this study we will first determine the rate of pathologic complete response to chemoradiation of ultrasound-stage II and III rectal cancers treated according to a standardized chemoradiation and surgery protocol, and evaluated by a systematic pathologic exam of the surgical specimen. We will then study whether complete pathologic response is equivalent to permanent tumor eradication. Patients with ultrasound-stage II and III rectal cancer treated with preoperative chemoradiation will be entered in a series of sequential Phase II trials in which the only variable will be the chemoradiation-to-surgery interval. A stable or increasing rate of pathologic complete response with longer chemoradiation-to-surgery intervals will indicate that pathologic complete response represents permanent tumor eradication rather than transient regression. In addition, we will explore the feasibility of using a panel of cancer-specific genetic alterations to detect cancer cells in the rectal cancer specimens of patients treated with preoperative chemoradiation and radical surgery. DNA extracted from cancer cells microdissected from pre-chemoradiation biopsy specimens will be tested for mutations in the p53, k-ras and APC genes, BAT26 deletions, and p16 promoter methylation. We will then investigate whether the specific mutation identified in the pre-chemoradiation cancer cells can be detected in a paired DNA sample obtained from the tumor bed and regional lymph nodes of the post-chemoradiation surgical specimen. This information will help develop molecular assays to diagnose pathologic complete response in transrectal biopsies of the tumor bed before removing the rectum.